Abstract
Automatic inspection processes require intelligent devices which combine the attributes of existing measuring techniques with those of the experienced practitioner. With the rapid growth of solid state electronics and computer applications during the last two decades, the field of computer vision has developed significantly and its application in the field of inspection is becoming a reality. This paper examines the use of computer vision for the inspection of engineering components. A range of existing edge detection algorithms has been tested for their suitabiliy for two-dimensional component inspection together with a new algorithm for application when contour tracing methods fail to detect localized edges in a complex scene.
Assessments are mainly subjective but are also based on computational speed; an important factor in on-line inspection applications. An algorithm has been proposed which can be used for the simultaneous analysis of internal and external contours with the ultimate aim of automatic component classification. In addition an alternative boundary coding system eliminates errors which are normally associated with other coding methods when monitoring components at different orientations.